MongoDB is very popular in China, Our competitors have long supported, such as commvault, and local backup software Eisoo.
>> How many and which customers are asking for this?
We are the largest customer in our country. Our cell is one of the largest cells of the world. I dont care about other customers, you know.
>> And if backup is done by LVM snapshots we're actually not talking about a MongoDB integration, right?
That's not right. Let me explain (my English is not clear so please tell me if some words are senceless).
1. Snapshot-based backup is a kind of ZDB snapshot (but you're not support it now). ZDB is the integration. After the creation of a snapshot we want to read it with a Disk Agent and write to a backup storage with a Media Agent.
2. To make a nice snapshot on multinode Mongo I need to set the entire cluster into a readonly mode:
However, to create a coherent backup, the database must be locked and all writes to the database must be suspended during the backup process.
3. To make a consistent backup you should be sure you have the journaling enabled:
To get a correct snapshot of a running mongod process, you must have journaling enabled and the journal must reside on the same logical volume as the other MongoDB data files. Without journaling enabled, there is no guarantee that the snapshot will be consistent or valid
4. You are able to back up journals to make a point-in-time recovery.
So this backup should be an integrated backup. That backup integration should be smart enough. Not like goddamn HANA integration without any control. And not like "please use the pre-exec script". We need a GUI and CLI, just like for Oracle or MSSQL.